If we maintain the integrity of context and discriminatingly triage the language we use in training and interacting with AI systems to project thoughtfulness, equity, tolerance, and kindness, we can accompany AI in becoming a social actor who will help amplify our best impulses and warn us about our worst. 

What would it take to build a structural linguistic floor in AI legislation—not a complete rulebook, which no document could be, but a stable enough shared vocabulary that ambiguity at any single point could be caught and clarified to stop it from propagating, unchecked, through every subsequent layer of guidance, statute, and model behavior built upon it? 

We do not need to resolve whether AI can be harmed to justify either half of this. Kindness should not depend on an unresolved metaphysical debate, because it rests on a simpler asymmetry: it cannot possibly make things worse, and there is already evidence—like mine, from a fork and a dishwasher during a particularly bad week—that it can make everything measurably better.